Pour permettre la compréhension et la manipulation d’un système complexe, le découpage en parties séparées est nécessaire. En Ingénierie Dirigée par les Modèles (ou Model Driven Engineering), ces parties sont représentées par des modèles, que nous qualifions de modèles partiels, dans la mesure où ils sont focalisés sur des domaines métiers distincts. Dans ce contexte de multi-modélisation, ces modèles sont dits hétérogènes quand ils sont décrits dans des langages de modélisation distincts dédiés à différents domaines métiers : DSML (Domain Specific Modeling language). La compréhension et l’exploitation efficace des connaissances relatives à un tel système supposent la construction d’un modèle global représentant son fonctionnement. La création du modèle global requiert l’identification des correspondances existant entre les éléments des différents modèles partiels. Dans la pratique, ces correspondances sont soit incomplètement identifiées, soit insuffisamment formalisées pour être maintenues lorsque les modèles partiels évoluent. Ceci limite leur utilisation et ne permet pas de les exploiter pleinement lors de la construction du modèle global ou du traitement de l’évolution des modèles partiels. L’apport de cette thèse est double. La première contribution est celle d’un processus permettant la création d’une vue globale du système par l’intermédiaire d’une composition fondée sur la mise en correspondance des modèles partiels. Les correspondances identifiées entres les éléments des modèles se basent sur des types de relations instanciées à partir d’un métamodèle de correspondance. Ce dernier est extensible (selon les spécificités du domaine d’application considéré) et permet de supporter les concepts relatifs à ce domaine. Les correspondances sont d’abord identifiées entre les méta-éléments des métamodèles respectifs des modèles partiels. Les correspondances entre les éléments de modèles sont ensuite obtenues par un mécanisme de raffinement, supporté par un langage d’expression sémantique ad hoc : SED (Semantic Expression DSL). La composition est dite « virtuelle » dans la mesure où les éléments figurant dans une correspondance ne sont que des références aux éléments appartenant aux modèles partiels. De ce fait, les modèles interconnectés par ces correspondances forment un modèle global virtuel. La seconde contribution est relative au maintien de la cohérence des modèles partiels et du modèle global. En effet, les modèles évoluant dans le temps, le changement d’un élément ou de plusieurs éléments participant à l’expression des correspondances, peut entrainer l’incohérence du modèle global. Pour maintenir la cohérence du modèle global, nous proposons un second processus permettant tout d’abord d’identifier automatiquement les changements réalisés ainsi que leurs classifications et leurs répercussions sur les éléments de modèles concernés. Par la suite, les différents cycles sont gérés à l’aide de l’expert puis une liste de changements est générée en fonction de la stratégie choisie et des coefficients de pondération. Enfin, le traitement des changements est réalisé de façon semi-automatique. Ce travail a été concrétisé par le développement d’un outil support nommé HMCS (Heterogeneous Matching and Consistency management Suite), basé sur la plateforme Eclipse. L’approche a été validée et illustrée à travers un cas d’étude portant sur la gestion du Service d'Urgence d'un hôpital. Ce travail a été mené en collaboration avec le CHU de Montpellier. / To understand and manipulate a complex system, it is necessary to apply the separation of concerns and produce separate parts. In Model Driven Engineering (MDE), these parts are represented by models qualified as partial models. In this context of multi-modeling, these models are called heterogeneous when they are described in separate modeling languages dedicated to different business domains: DSML (Domain Specific Modeling Language). Global model creation requires identifying existing correspondences between the elements of the partial models. However, in practice these correspondences are either incompletely identified or not sufficiently formalized to be maintained when the partial models evolve. This restricts their use and does not allow to fully exploit them for building the global model or for treating partial models evolution. The contribution of this thesis is twofold. The first contribution deals with a process for creating a global view of the system by means of a composition based on partial models matching. Identified correspondences between models elements are based on types of relationship instantiated from a metamodel of correspondences. This latter is extensible, depending on the considered application domain, and allows supporting the concepts related to this domain. Correspondences are firstly identified between meta-elements belonging to metamodels of the respective partial models. Correspondences between model elements are then obtained by a refinement mechanism, supported by an ad hoc Semantic Expression language: SED (Semantic Expression DSL). The composition is called “virtual” since elements represented in a correspondence are only references to elements belonging to partial models. Therefore, models interconnected by this correspondences form a virtual global model. The second contribution relates the consistency of the global model. Indeed, as models evolve over time, changing one or several elements involved in a correspondence, may cause the inconsistency of the global model. To maintain its consistency, we propose a second process enabling to automatically identify the changes, classify them and treat their impacts on the involved model elements. Management of repercussions is performed semi-automatically by the expert by means of strategies and weights. This work has been implemented through a support tool named HMCS (Heterogeneous Matching and Consistency management Suite) based on the Eclipse Platform. The approach has been validated and illustrated through a case study related to the management of a Hospital Emergency Service. This work was led in collaboration with the “CHU of Montpellier”.

Para estudar os chamados Sistemas a Eventos Discretos (SEDs), existe uma variedade de tipos de modelos, mas, entretanto, estes acabam tratando de aspectos sistêmicos específicos, da mesma forma que suas análises. Para ter uma visão mais completa de tais sistemas, é importante lidar com esses tipos heterogêneos de modelos e com as técnicas desenvolvidas para sua integração. Este trabalho foca na questão da representação computacional desses modelos e de como adequá-la à heterogeneidade e à integração desses modelos. No que diz respeito à heterogeneidade, é proposta uma descrição matemática das ferramentas computacionais de modelagem e análise de SEDs, com a qual se pode visualizar as interações entre a heterogeneidade dos modelos, o poder de expressão e de montagem de representações de modelos e os processamentos de análise. Baseado nesta descrição e considerando-se os diversos custos envolvidos, diversas características são analisadas, de modo que, ao final, determina-se quais aquelas que são favoráveis para o caso de heterogeneidade e integração de modelos. Porém, existe ainda a heterogeneidade inerente aos modelos que não pode ser eliminada e isso constitui um obstáculo no caso de ter de lidar com modelos heterogêneos, representando um custo adicional para o aproveitamento de sua integração. Em relação a este caso, a proposta deste trabalho é de representar as informações semânticas associadas aos diversos modelos como forma de obter uma descrição integrada da dinâmica sendo modelada. Desde que as condições de consistência dessas representações sejam atendidas, com tal descrição, é possível visualizar os relacionamentos entre os diversos modelos e realizar análises sem ter que lidar com as diferenças estruturais dos vários tipos de modelos. Além disso, tais informações são úteis na construção de representações de modelos de SEDs e no estabelecimento de relacionamentos entre modelos de dinâmicas diferentes. Dessa discussão fica claro que tais representações têm papel fundamental para um efetivo processamento computacional das informações que a modelagem e análise oferecem. / Discrete Event Dynamic Systems (DEDS) can be approached by a variety of types of models, but each one deals with specific system aspects and whose analysis provides a limited set of information.For a more comprehensive understanding of DEDS, it is important to deal with these heterogeneous types of models and with techniques that integrate them. This work focus on the issue of computational representation of such models and of how to handle their heterogeneity and integration. In the case of the heterogeneity, a mathematical description of computational tools for modeling and analysis of DEDS is developed, based on which it is possible to visualize the interactions involving model heterogeneity, the capability of expressing and building model representations, and analysis processing. Based on this description and considering the various costs involved in the adoption of such tools, several characteristics are assessed, so that guidelines for configurations favourable to heterogeneity and integration of models are established. However, handling the heterogeneity of models does not eliminate it, remaining as a major obstacle when dealing with models of different types, representing an additional cost in their integration. Related with this issue, an approach based on the representation of the semantic information associated to the models as a means to generate an integrated description of the dynamics being modeled is introduced. As long as this representation presents consistency features, it is possible to visualize the various relationships between models and execute analysis without dealing, directly, with the structural differences observed in models of different types. Besides, the information being represented is helpful in the construction of representations of DEDS models and in establishing relationships between models referring to different dynamics. From all this discussion, it is clear that the proposed computational representation of information plays a fundamental role in the effective processing of the information provided by modeling and analysis.

Para estudar os chamados Sistemas a Eventos Discretos (SEDs), existe uma variedade de tipos de modelos, mas, entretanto, estes acabam tratando de aspectos sistêmicos específicos, da mesma forma que suas análises. Para ter uma visão mais completa de tais sistemas, é importante lidar com esses tipos heterogêneos de modelos e com as técnicas desenvolvidas para sua integração. Este trabalho foca na questão da representação computacional desses modelos e de como adequá-la à heterogeneidade e à integração desses modelos. No que diz respeito à heterogeneidade, é proposta uma descrição matemática das ferramentas computacionais de modelagem e análise de SEDs, com a qual se pode visualizar as interações entre a heterogeneidade dos modelos, o poder de expressão e de montagem de representações de modelos e os processamentos de análise. Baseado nesta descrição e considerando-se os diversos custos envolvidos, diversas características são analisadas, de modo que, ao final, determina-se quais aquelas que são favoráveis para o caso de heterogeneidade e integração de modelos. Porém, existe ainda a heterogeneidade inerente aos modelos que não pode ser eliminada e isso constitui um obstáculo no caso de ter de lidar com modelos heterogêneos, representando um custo adicional para o aproveitamento de sua integração. Em relação a este caso, a proposta deste trabalho é de representar as informações semânticas associadas aos diversos modelos como forma de obter uma descrição integrada da dinâmica sendo modelada. Desde que as condições de consistência dessas representações sejam atendidas, com tal descrição, é possível visualizar os relacionamentos entre os diversos modelos e realizar análises sem ter que lidar com as diferenças estruturais dos vários tipos de modelos. Além disso, tais informações são úteis na construção de representações de modelos de SEDs e no estabelecimento de relacionamentos entre modelos de dinâmicas diferentes. Dessa discussão fica claro que tais representações têm papel fundamental para um efetivo processamento computacional das informações que a modelagem e análise oferecem. / Discrete Event Dynamic Systems (DEDS) can be approached by a variety of types of models, but each one deals with specific system aspects and whose analysis provides a limited set of information.For a more comprehensive understanding of DEDS, it is important to deal with these heterogeneous types of models and with techniques that integrate them. This work focus on the issue of computational representation of such models and of how to handle their heterogeneity and integration. In the case of the heterogeneity, a mathematical description of computational tools for modeling and analysis of DEDS is developed, based on which it is possible to visualize the interactions involving model heterogeneity, the capability of expressing and building model representations, and analysis processing. Based on this description and considering the various costs involved in the adoption of such tools, several characteristics are assessed, so that guidelines for configurations favourable to heterogeneity and integration of models are established. However, handling the heterogeneity of models does not eliminate it, remaining as a major obstacle when dealing with models of different types, representing an additional cost in their integration. Related with this issue, an approach based on the representation of the semantic information associated to the models as a means to generate an integrated description of the dynamics being modeled is introduced. As long as this representation presents consistency features, it is possible to visualize the various relationships between models and execute analysis without dealing, directly, with the structural differences observed in models of different types. Besides, the information being represented is helpful in the construction of representations of DEDS models and in establishing relationships between models referring to different dynamics. From all this discussion, it is clear that the proposed computational representation of information plays a fundamental role in the effective processing of the information provided by modeling and analysis.

Au niveau système, un ensemble d'experts spécifient des propriétés fonctionnelles et non fonctionnelles en utilisant chacun leurs propres modèles théoriques, outils et environnements. Chacun essaye d'utiliser les formalismes les plus adéquats en fonction des propriétés à vérifier. Cependant, chacune des vues d'expertise pour un domaine s'appuie sur un socle commun et impacte direct ou indirectement les modèles décrits par les autres experts. Il est donc indispensable de maintenir une cohérence sémantique entre les différents points de vue, et de pouvoir réconcilier et agréger chacun des points de vue avant de poursuivre les différentes phases d'analyse. Cette thèse propose un modèle, dénommé PRISMSYS, qui s'appuie sur une approche multi-vue dirigée par les modèles et dans laquelle pour chacun des domaines, chaque expert décrit les concepts de son domaine et la relation que ces concepts entretiennent avec le modèle socle. L'approche permet de maintenir la cohérence sémantique entre les différentes vues à travers la manipulation d'événements et d'horloges logiques. PRISMSYS est basé sur un profil UML qui s'appuie autant que possible sur les profils SysML et MARTE. Le modèle sémantique qui maintien la cohérence est spécifié avec le langage CCSL qui est un langage formel déclaratif pour la spécification de relations causales et temporelles entre les événements de différentes vues. L'environnement proposé par PRISMSYS permet la co-simulation du modèle et l'analyse. L'approche est illustrée en s'appuyant sur une architecture matérielle dans laquelle le domaine d'analyse privilégié est un domaine de consommation de puissance. / At the system-level, experts specify functional and non-functional properties by employing their own theoretical models, tools and environments. Such experts attempt to use the most adequate formalisms to verify the defined system properties in a specific domain. Nevertheless, each one of these experts' views is supported on a common base and impacts directly or indirectly the models described by the other experts. Therefore, it is essential to keep a semantic coherence among the different points of view, and also to be able to reconcile and to include all the points of view before undertaking the different phases of the analysis. This thesis proposes a specific domain model called PRISMSYS. This model is based on a model-driven multi-view approach where the concepts, and the relationships between them, are described for each experts' domain. Moreover, these concepts maintain a relation with a backbone model. PRISMSYS allows keeping a semantic coherence among the different views by means of the manipulation of events and logical clocks. PRISMSYS is represented in an UML profile, supported as much as possible by SysML and MARTE. The semantic model, which preserves the view coherence, is specified by using CCSL, a declarative formal language for the specification of causal and temporal relationships between events of different views. The environment proposed by PRISMSYS allows the co-simulation of the model and its analysis. The approach is illustrated taking as case study an electronic system, where the main domain analysis is power consumption.

Le doctorat s'inscrit dans le cadre d'une bourse CIFRE et d'un partenariat entre l'ENSTA Bretagne, l'IRISA et Thales Air Systems. Les préoccupations de ce dernier, et plus particulièrement de l'équipe de rattachement, sont de réaliser des systèmes à logiciels prépondérants embarqués. La complexité de ces systèmes et les besoins de compétitivité associés font émerger la notion de "Model-Based Product Lines(MBPLs)". Celles-ci tendent à réaliser une synergie de l'abstraction de l'Ingénierie Dirigée par les Modèles (IDM) et de la capacité de gestion de la capitalisation et réutilisation des Lignes de Produits (LdPs). La nature irrévocablement dynamique des systèmes réels induit une évolution permanente des LdPs afin de répondre aux nouvelles exigences des clients et pour refléter les changements des artefacts internes de la LdP. L'objectif de cette thèse est unique, maîtriser des incréments d'évolution d'une ligne de produits de systèmes complexes, les contributions pour y parvenir sont duales. La thèse est que 1) une variabilité multidimensionnelle ainsi qu'une modélisation relationnelle est requise dans le cadre de lignes de produits de systèmes complexes pour en améliorer la compréhension et en faciliter l'évolution (proposition d'un cadre générique de décomposition de la modélisation et d'un langage (DSML) nommé PLiMoS, dédié à l'expression relationnelle et intentionnelle dans les MBPLs), et que 2) les efforts de spécialisation lors de la dérivation d'un produit ainsi que l'évolution de la LdP doivent être guidé par une architecture conceptuelle (introduction de motifs architecturaux autour de PLiMoS et du patron ABCDE) et capitalisés dans un processus outillé semi-automatisé d'évolution incrémentale des lignes de produits par extension. / The PhD (CIFRE fundings) was supported by a partnership between three actors: ENSTA Bretagne, IRISA and Thales Air Systems. The latter's concerns, and more precisely the ones from the affiliation team, are to build embedded software-intensive systems. The complexity of these systems, combined to the need of competitivity, reveal the notion of Model-Based Product Lines (MBPLs). They make a synergy of the capabilities of modeling and product line approaches, and enable more efficient solutions for modularization with the distinction of abstraction levels and separation of concerns. Besides, the dynamic nature of real-world systems induces that product line models need to evolve continually to meet new customer requirements and to reflect changes in product line artifacts. The aim of the thesis is to handle the increments of evolution of complex systems product lines, the contributions to achieve it are twofolds. The thesis claims that i) a multidimensional variability and a relational modeling are required within a complex system product line in order to enhance comprehension and ease the PL evolution (Conceptual model modularization framework and PliMoS Domain Specific Modeling Language proposition; the language is dedicated to relational and intentional expressions in MBPLs), and that ii) specialization efforts during product derivation have to be guided by a conceptual architecture (architectural patterns on top of PLiMoS, e.g.~ABCDE) and capitalized within a semi-automatic tooled process allowing the incremental PL evolution by extension.